Bayesian Estimation of the Size of a Street-Dwelling Homeless Population

Abstract

A novel Bayesian technique is proposed to calculate 95% interval estimates for the size of the homeless population in the city of Edmonton using plant-capture data from Toronto, Canada. The probabilities of capture in Edmonton and Toronto are modeled as exchangeable in a hierarchical Bayesian model, and Markov chain Monte Carlo is used to sample from the posterior distribution. Guidelines are recommended for applying the method to assess the accuracy of homeless counts in other cities.